HAL Id: halshs-02925527https://halshs.archives-ouvertes.fr/halshs-02925527

Preprint submitted on 30 Aug 2020

HAL is a multi-disciplinary open accessarchive for the deposit and dissemination of sci-entific research documents, whether they are pub-lished or not. The documents may come fromteaching and research institutions in France orabroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, estdestinée au dépôt et à la diffusion de documentsscientifiques de niveau recherche, publiés ou non,émanant des établissements d’enseignement et derecherche français ou étrangers, des laboratoirespublics ou privés.

How to write Shertukpen (Sherdukpen) Language ?François Jacquesson

To cite this version:

François Jacquesson. How to write Shertukpen (Sherdukpen) Language ?. 2018. �halshs-02925527�

1

How to write Shertukpen (Sherdukpen) Language ?

François Jacquesson

These pages are an abridged version of the first parts of my book (2015): An Introduction to Sherdukpen

Language.

Those who are interested to know

which letters I used to write the language will go directly to section 1.2.

Those who want to know for which reasons I selected these letters (and relevant sound units, or

phonemes), are welcome to read part 2.

Languages are not pronounced in the same way among all speakers. This is also true for Shertukpen.

The language described here, is the language used, mostly, among the Thong clans in Rupa, with whom

I was mainly in contact. Descriptions of other varieties are of course welcome.

1. Introduction p.2

1.1. General context 2

1.2. Writing Sherdukpen 4

1.3. Word analysis 5

1.4. The technical lexicon 6

1.5. Abbreviations ans glosses 9

2. Phonology p. 12

2.1. Phonemes’ lists and minimal pairs 12

2.2. Vowels 17

2.3. Consonants 22

2.4. Syllables and word analysis 25

2.5. Morpho-phonemics and consonant clusters 29

2.6. Morpho-phonemics smaller points 33

2.7. Pronunciation and society 34

2.8. Quick speech 35

2

1. Introduction

1.1. General context

Sherdukpen language, as defined below, is spoken by c. 4000-5000 persons in the western part

of Arunachal Pradesh (in West Kameng district), in the North-East corner of India. Bhutan is close by

(30 km) to the west, and the Chinese (Tibetan) border is 70 km north as the crow flies, about 250 km

by road, after passing the great monastery of Tawang. Access to the rest of India is southwards, to

Assam, the capital city of which is reachable in 8 or 9 hours with a good car. Sherdukpens had, for a

long time and until 30-40 years ago, a regular three month winter stay in a host region of Assam; they

packed everything, gathered cattle, left their villages, formed a long file and journeyed southwards for

two or three days; this was called besme. Winter was passed down the hills, in Assam. When spring

approached, bamboo huts were abandonned, the file was formed again but upwards, and the

population came back to the villages - between 1500 and 2000 m. high. Many people still remember

that stage of the year. They were children or youngsters at that time, and their memories of those

winters are very lively and happy.

Map 2. The Sherdukpen country (Doimara is not indicated). Sherdukpen villages are

shown by dots. Domkho and Kelong are ‘Monpa’ localities; Wangho is a Bugun village.

The word ‘Sherdukpen’ is a recent coinage, which the British government of India supported,

and which is widely adopted locally, with two provisos. First, literate people prefer to write Shertukpen,

or Sertukpen, closer to the actual pronunciation. The name reflects the two components of the

community. The main group is the Tukpen (in the nearby Monpa language ‘Tukpon’), living in the little

town of Rupa (Thõ thük - thük means ‘village’) and the surrounding ten villages, the bigger ones being

Jigaon (Zagang) and Thungre. The smaller component, a bit farther west towards Bhutan, is Shergaon

(Sẽ thük), an isolated big village which everybody considers as founded by the main core of Rupa

3

people, in a remote past. The two words, Sẽ and Tukpen, or Ser and Tukpen, give Sertukpen, adapted

as Sherdukpen.

The language of Rupa and the surrounding villages can be considered as ‘Standard

Sherdukpen’, although the real Standard is now the language of Rupa town, which has lost its tones.

The language is still very much spoken, and most children do speak it fluently. We will see later (7.3)

what problems beset the transmission of the language. The language is not written, except in a few

official documench printed by the Tukpen Village Council (TVC), which acts as the official Secretariat

for the Village local authorities; and by some literate people, who try to adapt an English writing of

Hindi.

The Shergaon people, the Sẽ-ji, speak a very close dialect, and both groups can understand

each other without any difficulty. Although the Tukpen Council (as its name indicates) does not

interfere in Shergaon affairs, inter-marriages are frequent. The Shergaon speech (Sẽ-ji nyuk - nyuk

means ‘speech’) is considered very gentle, “with a nice tune”, by Rupa people who often consider their

own speech as “mixed”.

It is true that half the population of Rupa is “foreign”, mostly Nepali workers and Adivasi

servants (Adivasi here means people from the Tea Gradens1). Moreover, Indian Army camps are settled

in very many places around Rupa – but relations between Army personel and Sherdukpens, apart from

Canteen shops, do not seem to be very close. Hindi is widespread indeed, and this language often

appears in Sherdukpen conversations, except among older people. The more so among people who

have followed a longer schooling, but this category often speaks English as well. Moreover, the

generation who is now 50 or more often knows Assamese, because schools in this part of Arunachal

first were in Assamese, before shifting to Hindi or English. There is no local newspaper or radio: Rupa

is too small. Indian TV is everywhere, and Indian or English-language clips are watched on mobile

phones.

Rupa developped considerably since Independance, as photos by Verrier Elwin and others

show very well. In the 1950s, there were about 40-50 houses in Rupa, surrounded by jungle and

‘cactus’. The very few old beautiful stone-wood-bamboo houses that are left now amid the small world

of concrete are a drawning testimony of the Old Rupa. The great expansion came with the timber

business, during the 1970s and 1980s, an indirect consequence of the 1962 Chinese Aggression, after

which the Indian Army built roads to Assam. Big money was made at that time, especially by families

who owned the land, i.e. mostly Thong families. In 1993, the Supreme Court of India decided to stop a

business that had dramatically transformed the landscape and eroded the slopes. These were dark

times for the newly rich families, and stories run of ex-prosperous Thongs who secretly laboured at

night, in order not to be seen during the infamous act of working. Since 2000, an entrepeneurial re-

birth came with cash-crops, tomatoes and kiwis, and this benefitted a wider group of people. A major

source of big money is ‘contracts’, mainly in the road/bridge building etc. either for the Indian State or

the Army, as elsewhere in India. Yet prosperity, at least for the well-to-do families, is still linked with

land ownership, a hot topic, and contributes keeping most Chao families away from benefits. This

unequally spread prosperity has brought a large group of “foreign” workers in Rupa and around, as I

said before.

1 These people’s ancestors, mostly Munda language speakers, had been brought here from Orissa and neighbor

states by the British planters, to work in the thriving tea plantations. Most of them lived a rather secluded life

and hardly mixed with the local population.

4

There are a number of small villages around Rupa, all of them now reachable by more or less

mortorable road. Two of them are formed of Bhutanese people recently admitted (1990) into the

lower clans of Sherdukpen society; they still speak ‘Monpa’ as well as Sherdukpen. The other villages

have a longer pedigree of Sherdukpenship. All are included into the local tax system (chhkhok). The

bigger and most important one is Thungre (or Thongre), an old village where many of the rich families

of Rupa have lands and friends. The difference between centre (Rupa) and surrounding villages also

partly reflects the traditional and still perceptible cleavage between the Thong higher clans (Thong)

living now mainly in Rupa, Jigaon and Shergaon, and the lower Chao ones (Chhao) who live partly in

Rupa but mainly in the small villages.

Another Sherdukpen dialect (Sartang) is spoken, after a break in continuity, to the north,

mostly in the three villages of Rahung2, But (now Jirigaon) and Khoina. This dialect is distinct and to

some extant difficult for Tukpen people. It is possible to hear Tukpen and Sartang people shifting to

easier Hindi when speaking with each other. However, the closeness in speech is obvious to any

speaker of Standard Sherdukpen who happens to visit these places – but there are not many visitors.

These villages have long been considered ‘Monpa’3, and they are now politically considered as Sartang,

a specific category.

The higher clans of Sherdukpens, the Thong clans, agree that their ancestors came from far places in

Tibet, according to a “migration model” that is frequent in North-East India and elsewhere. The

founder-hero, Asu Japtong (Asu Gyaptong), is supposed to have come through these Sartang villages

with his warrior band. In some places, you may be shown stones or trenches that are supposed to be

related to his stay there. Priests (zizi) from these Sartang villages were for a long time considered as

the best specialists in ritual, and were regularly invited to perform at ceremonies in Rupa. They were

considered powerful and dangerous. In this scenario, the lower clans of present-day Sherdukpens, the

Chhao clans or some among them, are supposed to have had their ancestors in situ, and to be the first

inhabitants of Sherdukpen country. We will give indications about Shergaon and Rahung/Khoina

speeches when possible. Some comparative comments will then be added - but this book is not about

historical linguistics.

1.2. Writing Sherdukpen

Writing Sherdukpen implies special care for the 13 or 14 vowels; this will be described more fully in

parts 1.1., 1.2., 1.3. below.

There are seven oral vowels: a, e, i, o, ö, u, ü.

All have a nasal counterpart: ã, ẽ, ĩ, õ, ȫ, ũ, ǖ.

Each nasal vowel is written like the corresponding oral one, plus a “tilde” above it.

as in:

a la, ‘leg’, aha ‘yellow’, ba ‘fire’

e be’e’ ‘no’, nese ‘paddy’

ẽ sẽ ‘iron’, ẽˀ-pa ‘know’, herẽ ‘ribs’

i nyi ‘who’, abi ‘grandmother’, ami ‘mother’

ĩ ihĩ ‘root, nerve’, chĩ-ba ‘make sound’

o zo’ ‘upper floor’, abo ‘father’, oho ‘blue’

2 Whose inhabitants are considered to have come from Khoina, with a strong addition of Monpa people. 3 A cover term for a number of different people living on the fringe of Tibet.

5

õ azõ’ ‘white’, hatõ ‘picnic’

ö ta khö’-pa ‘spit’

ȫ dȫ ‘demon’, chhȫ-ba ‘wash’, gȫ ‘monitor lizard’

u su ‘meat’, du ‘son’, abu ‘brother’

ũ achũ ‘black’, asũ ‘plain, flat’, jũ ‘house platform’

ü skü ‘now’, bü-ba ‘carry on back’, gü ‘wealth’

ǖ ǖ-ba ‘wipe’, hǖ-ba ‘drop smthg’, gǖ ‘leather’

The only real groups of vowels seem to be /ao/ and /iao/.

For consonants, we use 23 (groups of) letters:

b, ch, chh, d, g, h, j, k, kh, l, m, n, ng, p, ph, r, s, t, th, w, wh, y, z.

It is also possible to write ‘ts’ and ‘tsh’ instead of ‘ch’ and ‘chh’.

In order to make a difference between short (or ‘checked’) and longer vowels, we write the short

ones with an apostrophe:

long short

blood ha ha’ food, rice

water kho kho’ stick

get up yao- yao’- steal

monitor lizard gö gö’ rope bridge

The checked vowels are often pronounced slightly higher.

For details and explanations, see chapter 2. Phonology.

1.3. Word analysis

This grammar intends to explain how the language is built, how sentences and words are made. For

this reason, we divide long words into their component units, and try to give explanations about what

these units mean. For instance, instead of Gatamji ‘Khoina people’, we write Gatam-ji, making clear

that Gatam means one thing (‘Khoina’) and ji another (‘group, people’). Another example is dacha

‘don’t run!’, which we write da-cha because da- here gives the negative meaning, and cha ‘run’. In

order to illustrate facts and problems, we give numerous examples. For instance:

wa ong-ba dükhüng-go gu ram-ba-ũ I came after he had gone

s3 go-Pf after-Loc s1 come-Pf-Ps

The first line gives the Sherdukpen sentence, divided in units, then a translation. Under the Sherdukpen

sentence come ‘glosses’: indications about the meaning of each unit. Some glosses are clear, like ‘go’,

‘after’, ‘come’. Some others are about grammar, and are given in a shortened form, for instance ‘s3’

means 3rd person singular, Loc ‘locative ending’ (one indication of place or time), Pf ‘perfect’ because

the -ba ending means a state of things arrived at (and is often translatable as a present tense). A

6

complete list of these shortened glosses is given below under § 5. Glosses begin with a capital letter,

except indications of person4.

1.4. The technical lexicon and its conceptual basis

My opinion is that books should be readable, and that any superfluous show of technicalities is to be

thrown overboard. Yet, things that are different have to be indicated as such, and this often requires

names or labels. Since it is hardly convenient (for art or memory) for instance to label suffixes by

numbers only, tradition and convenience conspire to give them names that suggest a meaning, or

something like a meaning. So, a suffix -lo whose meaning, when added to a noun, is that some creature

comes from that place or thing, a suffix which will often be translated in English by ‘from’ – will not be

labelled ‘from’ but ‘ablative’. Why?

Two benefits. The obvious one is that ‘ablative’ is not English but common to all trained linguists, be

they Chinese, French or Mexican: it escapes the difficulties that you automatically meet with ‘from’

when such a meaning is not worded like in English, but in a quite different way. Let us give one example.

In the Iliad by Homer, written in a pretty old and mixed dialect of Greek, somebody tells about some

children and uses the phrase5 hoi hethen exegenonto gunaikôn te thnêtaôn ‘who were issued from him

and mortal women’. It combines two different ways of indicating ‘where-from’ (ablative). ‘From him’

is hethen: he ‘he’ and the postposition –then ‘from’. This is very much like in Sherdukpen, where you

add –lo after a noun. But ‘and from mortal women’ is ex- (prefixed to the verb genonto) gunaikôn te

thnêtaôn where the meaning ‘from’ is indicated by both the ex- and the –ôn at the end of noun

gunaikôn and adjective thnêtaôn, except that –ôn also indicates there are several of them. Now, this

is very different from Sherdukpen or English, and also from hethen just before. Several techniques are

used, in the same language, in the same phrase; one that you can see as more simple, the other one

as rather complicated. Since they have the same meaning, you have to indicate both techniques under

one designation only, and it looks highly commendable not to suggest one of them is ‘better’. So, one

will use a technical term, ‘ablative’.

There is another benefit. This term ‘ablative’ belongs to a group of technical terms describing

directional movements, among which ‘ablative’ is contrastive. It indicates movement-from, while

other terms indicate ‘movement-to’, ‘movement-in’ etc. Here, to explain those terms, we use ‘from’,

‘in’ etc. but many languages have more possibilities and it may become necessary to describe these

specific movements with a complete and refined description. This description cannot be displayed

complete every time you meet the techique that indicates that movement, and here again a technical

term is welcome.

That once said, we have to admit that there exists in language description a far more shady area. This

is when you have to use general concepts or ideas without being sure they fit the language you

describe. A good example is ‘time’. Some philosophers like Aristotle or Kant decided, with some

interesting reasons, that it is difficult or perhaps impossible to think of the world which we live in,

without two basic notions, time and space. Consequently, one is inclined to look into languages, to

discover how time and space are modelled or expressed by people. We have just suggested that ‘space’

is treated in different ways, but at least we can see where it is at hand, and how it is described. Since

we can agree that there is something like ‘space’, we can look for its expression in many languages,

4 Writing ‘S3’ instead of ‘s3’ for singular 3rd person would be consistent with the principle that glosses begin

with a capital, but S and P in grammatical glosses often mean ‘subject’ and ‘predicate’.

5 Iliad 20.305.

7

and compare them. It works, or to a point. We perhaps should not push it too far, yet there is a wide

area where a number of people agree about what ‘space’ is, and about the varieties of possibilities

languages provide for expressing ‘it’.

The problem looks harder with ‘time’. Most people who went to school would probably say, on

request, that time is concerned with past, present, future. We have regrets, and expectations;

tomorrow sounds different from yesterday. A number of landmarks, sun, moon and memories, help

us to ‘map’ time and part of this mapping – as the term ‘mapping’ suggests – is space-like. In that view,

time looks like a thin and stretched bit of space we can run along, at least by thinking. The difficulty is,

when we again look into languages to discover how ‘time’ is worded, that it is not very clear. A great

number of other thoughts seem to interfere. In English, for instance, there is a striking contrast

between ‘past’, usually (but not at all consistently) marked with –ed on verbs, and ‘future’ where you

have to add a verb will and, if you insist on good English, shall. It is not difficult to realize that ‘you will

write’ is a kind of special result of ‘you want to write’, which is confirmed by the polite way it is

expressed when I do it: the ‘will’ is replaced by the ‘wish’ suggested by the now rare verb shall or its

past, should. The result is that, in English, ‘past’ and ‘future’ are not symmetrical, as we see from the

way they are expressed. ‘Future’ needs care, a special extra-verb, or even two of them, while ‘past’

needs only a flat suffix, or is half hidden in the vowels like in give and gave. ‘Future’ is clean, with its

special auxiliary verb, but will write then borders on many other areas like would write, should write,

can write, must write, all kinds of important nuances to which future belongs, much more than it

borders ‘past’. Actually, the group of nuances makes a fine set, where the person acts, feels, interferes,

wishes, etc. And this makes a quite different category or a quite different concept from ‘past’.

Once doubt begins, it is hard to stop it. After all, ‘past’ is not at all like ‘future’. Past is finished, is (or

was) something real even if it was dreams: at least I did dream those dreams. ‘Future’ is very different

and even if I am a man of prudent expectations, they have that cotton-like resilience I cannot expect

from my memories! Then, can I reasonably admit that ‘past’ and ‘future’ can be put up together and

be combined in somehting that is ‘time’? It looks all very hazy. Even without being too much a

philosopher, one easily sees or feels how shaky is the combination. It does not seems so obvious, after

all, that ‘future’ becomes ‘present’, and then ‘past’. There is some very queer gastric or digestive

metaphor hiddenly working in this tube-like process!

Should we reject ‘time’ in linguistic description? We should. Shall we do it? Hum. Readers would usually

see what sort of things we mean when we write ‘time’, and how else should we indicate that kind of

concern? Should we devise another ‘mapping’ of life and languages? Certainly, but how? Obviously

‘time’ has the dark quality of a default concept: we would wish a better one and a more lucid way of

seeing things – except that we only have a wish on one side and ‘time’ on the other. Since we are

describing languages, by which people are supposed to express their views, we could try to devise

philosophical concepts improving on ‘time’, that would be based on this or that language, on

Sherdukpen for instance. But would it stick to Sherdukpen, and the more faithful to Sherdukpen

language forms it would be, the more obscure or misleading for other readers. Moreover, would it

stick so well to Sherdukpen? We just described what happened with English: can we say that the way

English language is shaped, ‘expresses’ the way English-speaking people think or feel? Certainly not!

First because we do not change our views so easily when we shift from one language to another: a

language is a set of constraints all right, but not so much… if it were, we would become very ill every

time we shift from one language to another! Also, English and all languages are not really built as a

house or a pyramid. We saw that for the ‘past’, English has a number of techniques, the -ed technique

with variants (wanted, but heard, did, thought) and many more tricky ones (set, cast; ran, gave, saw):

which one does ‘express’ English lore correctly? The question is absurd. A category of ‘time’ that would

8

stick to English, would stick to what exactly? The result of this last exploration is bewildering. It boils

down to a sad consequence: we use concepts we are unable to justify, but also unable to substitute.

The good question now seems to be another one: how can we understand anything, since we seem to

work with such spooky tools? Is it really because of ‘tradition’? Are we really so much pasted or glued

to old pages, that our wings are unable to make us escape from the cage? That would explain the bad

sides of tradition, but not how we can manage nevertheless, and still less why we can so clearly criticize

what we live by. The explanation is that terms do not come alone. For all its fuzziness, a term like ‘time’

is seen as one piece only in a larger game where, not only the neighbouring ‘space’, but their areas of

use (‘verbs’, ‘adverbs’, ‘adpositions’ etc.) efficiently contribute to clear up meanings. A term like ‘past’

is certainly debatable, as well as ‘present’, ‘future’ and many others – but taken together and used in

a description with examples, the reader soon realizes how their very imperfect frontiers come, more

or less, to make provisonal maps. If you use ‘future’ by itself, you never know what your reader may

understand, and you are lost. But for instance in a contents page, the reader may make one’s ideas

about how terms are juggled with each other, or set against each other – or the metaphor of N-

dimensional space you may find more familiar.

In a way, this makes Kant (or perhaps a simplified view of what he wrote) turned inside out. The

philosopher Kant researched what in the mind is indispensable for perceiving the world. Although he

admitted that the objects we perceive are the indipensable stimulus for thought, he reasoned that

they were not sufficient and could not explain alone how we think; there should be conceptions

somehow before these objects or at least different from them; these conceptions with the objects we

perceive would make intelligence possible. His favourite examples for helping us to realize that some

ideas do not depend on the objects we perceive, are ‘all changes must have a cause’ and the fact that,

according to him, when imagining an object we can dispense with colour or reflections, resilience or

weight, but not with the ‘space occupied by the object’; these two examples already suggest, if not

contain, his idea that basic thought cannot do without space and time. In Kant’s method, space and

time are what imagination or thought cannot do without; somehow they are what is left when all else

is relaxed. So, space and time appear as general conditions for the undertsanding of anything.

Yet, ‘space’ and ‘time’ are something different for the people who write grammars, and certainly for

people who read them. Here ‘space’ and ‘time’ operate as super-labels for describing specific things,

not all of them. ‘Space’ will be useful only if you contrast it with something else, for instance ‘agency’

or ‘gender’, because in a collection of available affixes you have to classify which is which. The same is

true with ‘time’, althought it often concerns verbs (but never only verbs) rather than nouns, because

for instance among the possible affixes to a verb, some may suggest something like time, but others

do suggest different criteria like person, number of agents of patients, reciprocity, negativity etc. In

such a descriptive and contrastive context, ‘space’ and ‘time’ are not categories for a general

understanding of things or events, but categories that help classify a group of forms into subsets.

The reason why we use terms like ‘space’ ans ‘time’ in grammatical description is then rather different

from what we mean by ‘space’ and ‘time’ in a more general manner. ‘Space’ in grammar has to do with

‘space’ in general, but relatively, only in that measure it helps sorting out things; this or that suffix or

grammatical pattern will be described under the label ‘space’ because it (rather) has to do with space,

not because it describes space nor really places things in space. This becomes obvious when we

compare languages. The ambition regarding space is different from language to language, and very

often, there are side effects with forms that are supposed to describe ‘space’. A typical example, also

valid for Sherdukpen, is with demonstrative pronouns or adjectives. Often, they are described as

‘closer’ or ‘farther’ to or from the speaker or his/her representative; this does suggest a description of

space. Yet, in most cases these demonstrative words do not tell if the designated thing or person is

9

closer or not, but just provide a listing effect, for instance when you ask ‘do you prefer this or that?’ a

question that has rarely to do with distance but with colour, form, taste or any contribution to choice.

Most notably, in English and often, ‘that’ is a neutral form and ‘this’ is a marked one; this is why it

makes (or may make) a difference if you ask ‘have you seen that?’ or ‘have you seen this?’, once again

without any distance ingredient. ‘Space’ is then a difficult category, and can rarely be taken at face

value; the same is true with time. The bigger the concept, the bigger the risk to have it used in an

unexpected way.

We could perhaps learn something for such an delusive principle. As within language, where words do

mean something, but do it all the more when working in the flow of speech or in the elaboration of

narrative, so without the language when we use words to pinpoint phenomena and when we try to

sort out ‘how it works’. Language description, either when we write it or when we read it, teaches us

a reasonable skepticism: ‘person’ does not mean ‘person’, ‘time’ does not mean ‘time’, ‘space’ does

not mean ‘space’, these words sound like masks in some Commedia dell’ Arte or other comedy, but

this is the firm principle of theater. A mask looks unconvincing when alone, yet things change when

the intrigue is running.

1.5. Abbreviations and glosses

Nearly all abbreviations are given in the two charts below. The first chart gives abbreviations in

alphabetical order; the second one gives words or suffixes in alphabetical order. Other abbreviations

are for personal pronouns: s1 s2 s3 denote 1st, 2nd and 3rd person singular; p1 p2 p3 for plural.

For an analytical list of grammatical suffixes, see chapter 5. The last column of the chart gives the

chapter and section where the grammar point is discussed.

1.5.1. Alphabetical list of abbreviations in glosses

A -‘o agent marker 2.1.1

Abl -la, -lo Ablative 2.2.3

Adl -ro Adlative (Khoina) 2.2.3

Adv -no Adversative topic 2.2.4

An -zing until now 2.6.10

Cnt -bo Continuative 2.6.6

Com -la Comitative 4.7.3

Cv -ro Converb 3.1

Dbt bam Doubt 2.6.2

Df d°-, n°- Defence (Ng Imperative) 1.2.4, 2.8

Dir -ta, -to Direction

Ds -rang Distributive indefinite 2.1.5

Du -zing Dual (with pronouns) 2.1.1

Ext -thẽ Extensive 2.5.8

F -mu feminine

Fac -chhi- Factitive 2.5.3

Fe -mat female on N

Ft -mu future 2.6.8

Gp -na Gap in time fut/past 2.6.9

Ind -nya.ĩ Indefinite pronoun 2.1.5

Int -ni, -ĩ Interrogative 2.2.3

Ip -mo Imperative 2.4.8

10

Ipf -ma Imperfective aspect 2.4.4, 2.6.5, 2.6.8

Ite -da- Iterative 2.5.6

Loc -go Locative 2.2.3

NCn -raĩ, -rẽ Negation of continuative

Ng b°-, m°- Negation 1.2.4, 2.8

NGp -do No gap in time fut/past 2.6.9

NGp-Fut -do-m 2.6.8

NgL bo’o negative locative predicate 2.8

NgN be’e negative predicate 2.8

NTm -la Non witness 2.6.5

O -ni, -ĩ patient marker 2.1.1.

Obl -che’ Obligative: ‘must’ 2.5.7

Obl-Fut -che’-ma 2.6.8

Pf -ba, -pa Perfective aspect 2.6.5, 2.6.7

Pfp -ba-ũ past tense 2.6.7

Po -bu Availability, possibility 2.5.9

Po -lan From now on 2.6.10, 2.5.10

Pos -ũ, -õ Possessive, genitive 2.2.1

Pros -ra’a Prosecutive 3.2

Ps -õ Past 2.6.7

PsR -nyi Past recent 2.6.7

Qn -a, -ã Question 2.1.4

Rec -di’ Reciprocal, reflexive 2.5.4/5

Ref -ji Reflexive 2.1.2

Sp -po Specifier, with nouns 2.6.1.

T -gi Topic 2.2.4, 2.6.1

1.5.2. Alphabetical list of affixes

-a, -ã Qn question mark 2.1.4

b°-, m°- Ng Negation 1.2.4, 2.8

-ba, -pa Pf Perfective aspect 2.6.5, 2.6.7

-ba-ũ Pfp past tense 2.6.7

bam Dbt Doubt 2.6.2

be’e NgN negative predicate 2.8

-bo Cnt Continuative 2.6.6

bo’o NgL negative locative predicate 2.8

-bu Po Availability, possibility 2.5.9

-bu for years past 4.2

-che’ Obl Obligative: ‘must’ 2.7

-che’-ma Obl-Ipf 2.7

-chĩ Abi Abilitative 2.7

-chhi Fc Factitive 2.7

-chho Cap Capacitative 2.7

d°-, n°- Df Defence (Ng Imperative) 1.2.4, 2.8

d°--ni for days past 4.2

-da- Ite Iterative 2.5.6

11

-di’ Rec Reciprocal, reflexive 2.5.4/5

-do NGp No gap in time fut/past 2.6.9

-do-m NGp-Ipf 2.6.8

-ji people 4.3.1

-ji Ref Reflexive 2.1.2

-gi T Topic 2.2.4, 2.6.1

-go Loc Locative 2.2.3

-la, -lo Abl Ablative 2.2.3

-la Com Comitative 4.7.3

-la NTm Non witness 2.6.5

-lan Po From now on 2.6.10, 2.5.10

lin- for days to come 4.2

-ma Ipf Imperfective aspect 2.4.4, 2.6.5, 2.6.8

-mat Fe female on N

min- for years to come 4.2

-mo Ip Imperative 2.4.8

-mu Ft future 2.6.8

-mu F feminine

-na Gp Gap in time fut/past 2.6.9

-na-m Gp-Ipf 2.6.8

-na, -ã Qn Question

-ni, -ĩ O patient marker 2.1.1.

-ni, -ĩ Int Interrogative 2.2.3

-no Adv Adversative topic 2.2.4

-nya.ĩ Ind Indefinite pronoun 2.1.5

-nyi, ĩ Pst Past recent 2.6.7

-õ Ps Past 2.6.7

-‘o A agent marker 2.1.1

-po Sp Specifier, with nouns 2.6.1.

-ra’a Pros Prosecutive 3.2

-raĩ, -rẽ NCn Negation of continuative 2.6.3

-rang Ds Distributive indefinite 2.1.5

-ro Adl Adlative (Khoina) 2.2.3

-ro Cv Converb 3.1

-ru Def

-ta, -to Dir Direction 2.1.6

-thẽ Ext Extensive 2.5.8

thi also 2.2.5

-ũ, -õ Pos Possessive, genitive 2.2.1

-zing Du Dual (with pronouns) 2.1.1

-zing An until now 2.6.10

12

2. Phonology

This chapter is devoted to the sounds of the language. Some of these sounds are found in many

languages, but some others are rare. We would like to praise the Sherdukpen people who suffered

repeated questions about sounds, when they were anxious to teach the meanings. We remember one

evening with Tsering and Khandu Thongdok, who gathered elderly experts around a bokhari, and I

spoiled the feast by harassing them with questions about rare sounds like /ẽ/, /ĩ/ and /ȫ/ (see

hereafter). They were amused. More recently, Dorje Khandu Thongon, with the help of his wife Pema

Chhom, took a careful interest in checking the accuracy of my chart of sounds, and the whole glossary

– a rare feat. Many people took care of repeating words and sentences several times: to all of them,

thank you.

2.1. Phoneme lists and minimal pairs

2.1.1. Initials and finals

Some sounds (phonemes) are not found, or rare, in the beginning of words, or are never found at the

end of syllables. It is then useful to make two lists.

1/ Initials

m b p ph w wh

n d t th l

g k kh y

ny z s r

j ch chh h

a e i o ö u ü

ẽ ĩ õ ǖ

2/ Finals

a e i o ö u ü

(ã) ẽ ĩ õ ȫ ũ ǖ

ao

aõ

m p

n t

ng k r ‘

Comapring these two lists, we can guess for instance that the ‘ng’ soung is rare or never at the

beginning of words, but can be found in the ending. On the contrary, ‘ny’ is found in the beginning, but

never at the end.

2.1.2. Minimal pairs

In order to make sure which sounds are different, you have to find examples that make the difference.

For instance lü and lu mean different things (respect. ‘body’ and ‘mountain’), then you may feel

confident that the sound /u/ and the sound /ü/ are not the same. Another example: cha! means ‘run!’

and chha! means ‘buy!’. It follows that /ch/ and /chh/ are significantly different sounds, and that you

must catch this difference if you speak Sherdukpen. Words in pairs like lu and lü, or cha and chha, are

13

called minimal pairs. Of course you can contrast more than two words, as we have done in the charts

below.

Vowels

In the first chart, we find 9 different words, all with sound /g/+ vowel. This helps showing that these 9

vowels are distinct. The same process applies in the following charts.

a ga we

ao gao reliquary necklace (< Tibetan)

i gi (topic in grammar)

o go saddle

ö gö rope & cane bridge

ȫ gȫ monitor lizard

u gu I

ü gü wealth, gü-ba ‘like, love’

ǖ gǖ-ba make faces, transform oneself

a ha blood; ha’ ‘food, rice’

e he’-pa spread (water, ash etc.)

ĩ hĩ flat land; hĩ-ba ‘getting well’

o ho’-pa pull out something, be hanging; chop

õ hõ-ba peel (skin)

ũ hũ’ salt

ǖ hǖ-ba drop, let drop

ao khao akhao ‘elder’, khao-ba ‘snatch’

e khe’-pa cry

i khi cane

o kho water

ö khö’-pa to spit

u khu five

ũ ba-khũ charcoal

ü ba-khü smoke

a la’ leg, foot

e lele’ beauty

ẽ slẽ spleen

i li’ bow

o lo south ; lo-ba ‘plant(seeds)’

u lu mountain

ü lü sacred stone & its inhabitant

o pho’ yeast

õ phõ flag

u phu mountain god

ũ phũ tiger

14

ü phü insect, phü-ba ‘present, offer’

a sa’ poison

ao sao hyena

e se rhino

ẽ sẽ iron

i si grass, fodder

ĩ sĩ-ba smell (bad)

o so’-pa speak

õ sõ ten

ȫ sȫ Sichuan pepper, ‘jabrang’.

u su’ meat

ũ sũ a draw-bridge

ü sü-ba meet

ǖ sǖ a little bit of something

ẽ thẽ- take

i thi this

o tho’ oil

õ thõ thük Rupa

ö thȫ bridge

ü thü that other side

ǖ thǖ thǖ-ba ‘cover with basket’

a za- laugh

ao zao- mix

e ze- eat (highly honorific)

i zi’ Fr. carquois

ĩ zĩ- lead (the way)

o zo’ upper floor

õ zõ- clear the guilt

ö zö- stay aside

ȫ zȫ: goat

Putting those charts together, we find:

a e ẽ i ĩ o õ ö ȫ u ũ ü ǖ ao

g + + + + + + + + +

h + + + + + + +

kh + + + + + + + +

la + + + + + + +

ph + + + + +

s + + + + + + + + + + + + +

th + + + + + + +

z + + + + + + + + +

15

All vowels that are found after the same consonant (in the same line above) are contrastive. For

instance, the fact that /u/ is different from /ü/ is clear because both occur after /l/ or /kh/ or /g/ or

/s/6.

If we look at the four /o/, /õ/, /ö/, /ȫ/, the difference

- between /o/ and /õ/ is proved because both occur after /kh/, /h/, /s/, /z/;

- between /o/ and /ö/, because both occur after /kh/, /g/, /z/;

- between /o/ and /ȫ/ because both occur after /s/, /z/;

- between /õ/ and /ö/ because both occur after /kh/, /z/;

- between /õ/ and /ȫ/ because both occur after /s/, /z/;

- between /ö/ and /ȫ/ because both occur after /z/.

In the same way, the chart helps checking which vowels can be proved distinct or not, from the lists

of words above. Before concluding, we have to take into accounch new lists.

Oral / nasal contrast

e ẽ

beauty lele slẽ spleen

rhino se sẽ iron

give birth ke-ba kẽ near

i ĩ

to die i-ba ĩ-ba to wipe

grass, fodder si sĩ-ba smell bad

o õ

to see o-ba õ’õ dry

blue oho dahõ vegetable (a species)

tax kho ba-khõ charcoal

(imperative suffix) mo mõ butter

be hanging (cloth) no-ba anõ-du slowly, softly

to speak so-ba sõ ten

to write zo-ba zõ-ba clear the guilt

ö ȫ

cane & rope bridge gö gȫ monitor lizard

u ũ

potato ju jũ platform bef. house

ü ǖ

spiritual son rübi rǖ thick jungle

wealth gü gǖ make faces, transform

meet; suffocate sü sǖ ait little

The ö / ẽ contrast

ẽ ö

near kẽ chökö’ bee (species of)

6 This is a shortcut. The meaning is: because both vowels occur after /l/ in words having distinct meanings, etc.

See above.

16

ẽ ȫ

liver ẽchhẽ dȫchhȫ’ wasp (a big species)

and

o-ba to see

ö’-pa to be in hurry

õ-ba to be dry

Several points remain unclear. Some will be settled in later sections.

Consonants

i-ba to die

b abi grand-mother

d di’-pa ik-khat di’-pa ‘arm wrestling’

j ji’ others

g gi (topic)

kh khi’ cane

l li bow

m ami mother

n nini sun

ny anyi female of other clan

r ri-ba burn

s si grass

t jati spear

th thi this

chh achhi father

y yi-ba distribute

z zi urine

Contrast for alveolars and affricates:

ch cha-ba to run

chh chha-ba to buy

z za-ba to laugh

s sa’ poison

n / ny

softly anõ-du anyõ proper, properly

shoes nunu nyu fish

(interrogation) ni nyi who?

Some pairs for the final check are (see 1.2.1.):

fire ba: ba’ present, be there

water kho: kho’ stick

blood ha: ha’ food, rice

get up yao: yao’ steal

dice su: su’ meat

17

monitor lizard gö: gö’ hanging bridge

nature’s call hũ: hũ’ salt

urine zi: zi’ Fr. carquois

A rare sound is /jy/,probably from /j/ + /iao/ as in jiao ‘thief’, different from diao ‘yesterday’.

2.2. The vowels

We will here describe some touchy points about vowels, and their groupings.

The full range of contrasting vowels is:

a e i o ö u ü

ã ẽ ĩ õ ȫ ũ ǖ

With one frequent diphthong /ao/ and a rare one /aõ/:

ao

aõ

2.2.1. Vowel length, pitch, tones

Tones still exist in three villages: Düksü, Mukhuting, Gorbao. They are also more or less audible among

older people elsewhere, but not systematically. They disappeared in the bigger villages like Jigaon or

even Thongre, and cannot be often heard in Rupa. The influence of Assamese and Hindi is probably

the cause for this transformation. Since this description is about Rupa’s speech, tones are not taken

into account. Rarely, you can find in Rupa words with an “abnormally long vowel”, for instance in alêdu

‘pretty’ or yî-di’-pa ‘distribute property’.

What can be heard systematically, at least among elder people, or in the speech of those among the

younger ones who have learnt the language carefully or with their grand-parents, is a difference

between vowels that are abruptly stopped, and vowels that are not. The easiest description is to say

that some vowels are ‘short’ and that others are ‘long’, but this does not convey the real sound. It is

more exact to say that some vowels sound ‘shortened’ or ‘checked’, while others are not. In this book,

the shortened ones are written with an apostrophe: tha’ is with a shortened vowel, while tha is not.

Easy examples are (see also: 1.1.2. Minimal pairs, 1.3. Consonants):

‘long’ ‘short’

fire ba ba’ present, be there

mother’s sister anyi anyi’ brains

reptile gö gö’ hanging bridge

food ha ha’ blood

nature’s call hũ hũ’ salt

du ke-ba give birth ke ke’- charge, burden

water kho kho’ stick

Assam plains nyu, nyũ nyu’ fish

be itching o- o’- kill

cattle, cow spu spu’ owner

dice su su’ meat

bridge thö, thẽ thö’- throw

get up yao yao’ steal

urine zi zi’ Fr. carquois

18

yack zizi zizi’ priest

The difference is best described in terms of length, but the short syllable can be heard as checked or

higher in tone. Speakers with whom we discussed the point agree that the ‘long one’ is heard like a

long and quiet vowel, while the other one is swifter and abruptly ended; the description by relative

height is not much convincing according to them. In this book, we wrote the short syllables with the

apostrophe. We write the long vowels with “:” only when necessary.

Other examples of words with /’/:

anyi’ brains khyi’- quake

be’e’ no le’- stop (liquid)

Bisri’ Bisiri (river) nyü’- put on (pants)

du’ son nö’- be affected

ji’ stranger o’- kill

ẽ’- identify pho’ yeast (rice beer)

go’o I (transitive) ra’- do, prepare

he’- spread (water) ro’- walk

hũ’ salt sa’ poison

khe’- cry whu’ bird

yo’ clothes

The consonant character of this phoneme is made obvious by its triggering the /pa/ variant of the verb

suffix instead of /ba/, as the final consonants /p, t, k/ do. For instance, we find the /ba/ variant with

cha-ba ‘run’, chi-ba ‘give’ or ching-ba ‘pluck’, but the voiceless /pa/ variant with chek-pa ‘cut’, rok-pa

‘hang’ and ro’-pa ‘walk’.

2.2.2. Vowel descriptions

Vowels /a/ and /o/

Probably because of Assamese influence, there is a non-phonemic indecision between /a/ and /o/ for

a good many words. Since Assamese people pronounce /ɔ/ “open o” the older “short a”, there is a

fashion for pronouncing in this way what are actually /a/ vowels. Examples are nyakha ‘where’,

commonly pronounced [nyɔkhɔ], or ta ‘towards’ pronounced [tɔ]. The frequent ablative suffix is mostly

heard -lo, but -la is quite possible. Normally, the Sherdukpen /o/ is pronounced [o] in closed syllables

and [ɔ] in open ones: the risk of confusion with ‘true o’ exists. This is a problem, since /a/ and /o/ are

distinct phonemes in the language: la ‘leg’ and lo ‘god’.

It then seems quite possible that a number of words in which we now hear /o/ were with /a/ in not

such a remote past. For instance, in the oldish phrase that follows, that has a decided proverb-like

flavour, ‘water’ is kha, not Standard Sherdukpen kho.

wũ-ngo kha-go in jungle and river

jungle-Loc river-Loc

Vowel /ü/ and /i/

The i / ü contrast

i ü

19

prayer nying nyüng breast, milk

(topic) gi gü wealth; to swallow

cane khi ba-khü smoke

bow li lü inhabitant of sacred stone

women coming from the

same clan

miring mürüng jackal, fox

sun nini nünü baby

NB : nunu ‘shoes’

The vowel /ü/ is frequent although in many cases younger generations pronounce /i/ instead. A word

like skü ‘now, presently’ (*ski is never heard) is frequent in the speech of everybody. A word like phü

‘insect’ (a better gloss is ‘small animal’ since spiders also are phü) can have a more central or less

rounded vowel, but never a decided [phi], as far as we know. Yet, in bi-syllabic nouns with /i/ and /ü/,

it may be difficult to discern the right sequence: /i-i/, /i-ü/, /ü-i/, /ü-ü/, and good speakers are required

to give their advice.

Vowels /ö/ and /ȫ/

bö ginger like plant

böchhö’ gun

gö, gẽ monitor lizard

gö’ hanging bridge

khö’-pa eject, spit; stear, mix

khrö-ba open (a box)

nö’-pa harm

ö-ba be in hurry

thö’-pa, the’-pa throw

chökö a wasp

chhö ram-ba wake up

Chhölö, Chhẽlẽ Chilipam

zö-ba stay aside

The Lexicon gives a rather small collection of /ö/. Younger speakers tend to pronounce /e/ instead of

/ö/, and /ẽ/ instead of /ȫ/, for instance in thö or nethö. The Sherdukpen autonym mȫ is now often

pronounced mẽ. When visiting the former settlement in Rochong, R. K. Karma described the ladder,

hing lȫthȫ, with which one ascended, in times past, the upper platforms where the youngsters passed

the night; D. K. Thongon re-pronounced it as hing lẽthẽ, but was corrected by his elder. Confusions

between /ö/ and /ẽ/ can also be heard, as the charts show.

The sound /ȫ/, the nasalized version of /ö/, is found in:

chhȫ-ba wash (cloth)

chhȫ-ba wake (somebody)

dȫ demon

dȫchhȫ a big wasp

hȫsu coriander

lȫthȫ, nȫthȫ ladder

mȫ Sherdukpen

sȫ, sȫĩ Sichuan pepper

thȫ, thẽ bridge

20

thȫ-ba sleep (for birds only)

zȫ goat

The final /ĩ/ which is heard in the ever-present sȫĩ (the Sichuan pepper that Assamese people call

jabrang, and that played an important role in the trade between Assam and Rupa) is a problem,

complicated by the relatively strong variation in the pronunciation of this word. It may be the result of

an older /sön/; see 1.1.3 § 3.

Other consonants, such as *r, also seem to have played a role, but this chapter refrains from

etymological discussion. The /ö/ is rare among younger speakers, who regularly substitute /ẽ/, and the

same is true with /ü/ often substituted by /i/.

Vowels /ĩ/

The existence of /ĩ/ as a phoneme is clear. Several examples follow /h/. Here is a list:

angyĩ ’ handle of knife

achhĩ sweet

hĩ flat land, plain

hĩ-ba fall down

ihĩ root, nerve

ĩ-ba wipe, clean

chĩ mucus

chĩ-ba to make sound

zĩ-ba lead

Vowel /ã/

The case of /ã/ is more embarrassing. It is heard in few cases, especially when it seems that /n/ or /ng/

has been lost. Examples are:

(1) lok-dão, a rapid-speech form, see /lok-dango/ in Jigaon;

(2) nyakhã or nyakha or nyakho ‘where?’;

(3) skü-ã, for skü-nã or skü-na ‘now’;

(4) khãi ‘20’ – a form that has a number of local variants, among which khan and khoan, the [n] and

nasalization of which might be explained by the Khoina form kho-han with han meaning ‘1’.

The aĩ form for han ‘one, 1’ is very widespread. Among numerous examples: alangaĩ ‘a little’ (alang

han), or jepoaĩ ‘much, a quantity’ (jepo-han).

2.2.3. Vowel groups

The /ai/ group is rarely found. The /ai/ in Ai-meme ‘grand-parents’ certainly is a short form for ayo,

one term for ‘grand-mother’. Baidong ‘singer ritual group’, although not a Tibetan word, is probably a

borrowed word. Jowai ‘100’ and khãi (and variants) ‘20’ are shortened forms, and would be more

appropriately written jo-aĩ and kha-aĩ or kho-aĩ, where aĩ < han ‘one’.

The /ao/ diphthong is common, and the /iao/ group is found in a few words. This /ao/ is not always a

straightforward diphthong. For instance, there is a phonetic difference between yao-ba ‘steal, rob’ and

yao-ba ‘wake up’. The first one is accented on [a], not the second, which should perhaps be written

/yawo-ba/. I know only one example of /aõ/: haõ ‘salty plot of land’.

21

A fex examples of /ao/:

achhao sharp

akhao eldest

chhao the ‘lower clans’

chhao’ nivrea, fish poison

dochhao mouth

khao door (house).

phao-ba to dry

sao-ba look for

tao-ba to climb

zao-ba to fry on fire

Examples of /iao/ are:

diao yesterday

jiao thief

khiao flying cat

miao chest

phiao knot

Other vowel groups beginning with /i/ do not seem to occur. But we have to take into account the /ny/

consonant. Maybe the origin of, for instance, diao is *di-yao; and we could imagine a similar historical

development for the few other cases; but diao etc. is a better writing for the actual pronunciation.

The /uo/ group is found in juõ ‘dumb, stupid’ and suo ‘ox’; it might be analyzed as /uwo/.

2.2.4. Words beginning with vowels

Except for /a/, a prefix with a grammatical role, vowels are not common at the beginning of words. In

our lexicon, the distribution is:

a 91 ã

e 2 ẽ 4

i 22 ĩ 1

o 17 õ 1

ö 1 ȫ

u 7 ũ

ü ǖ 1

A few observations are needed.

The number of words beginnings with /i/ is important only because of the numerous compounds with

ik ‘hand, arm’. If we do not include them, the count falls to 7.

The number of words beginnings with /o/ is reduced to 9, if we exclude the forms where /o/ is the

phonetic variation of /a/ under the influence of a following rounded vowel, for instance ojo (< ajo)

‘high’, oho (< aho) ‘blue’, ohomiya (from Assamese < ahomiya) ‘Assamese’, oyung (< ayung) ‘shadow’,

ochho (< achho) ‘on the right’, ozu, a technical variant of azu.

The same remark is valid with /u/. Words beginning with phonemic /u/ are uhu ‘ceremony’ (arguably),

uk-pa ‘hide’, ung ‘three’, ur-ba ‘provoke pain’.

22

In the Lexicon, we wrote some of these words two times, for instance ajo and ojo. So they are found

here in our ‘census’. The reason is, that although the /a/ variant exists, a proportion of people would

look for these words under the other vowel. Depending on local tradition, the proportion can be higher

or lower. We did not indicate the /a/ veriant when it is not heard, even if we could have reasons to

think it is etymological. A good example is ehek ‘red’, which comes from *a-hek but is never heard

*ahek.

Words beginning with a nasal vowel are rare. See ẽ-ba ‘ride’, ẽ’-pa ‘know’ (and its derivative ẽdi’pa), ĩ-

ba or ǖ-ba ‘wipe, clean’ and õ’õ ‘dry’. Note that all are or are derived from mono-phonemic verbs.

2.3. The consonants

2.3.1. Nasal initials

The four points of articulation (m, n, ny, ng) are clearly distinguished, although the existence of ng- as

an initial is disputable and (as far as this research could reach) is found only in ngüng-lingma-thok, the

name of a creeper. Its appearance can be due to ‘Monpa’ or Tibetan influence: the language spoken in

Kalaktang and Dirang has words with ng- in the beginning, as in Tibetan. In Standard Sherdukpen, we

have /ng/ between vowels in very few cases: in lok-dango ‘bag’ (in Jigaon), bẽngo ‘deaf’, and last but

not least thongõ - the name of one of the most influential clans in Shedukpen country, although the

traditional writing as ‘Thongon’ has induced a [thong-gon] pronunciation, criticized by many7; it is not

difficult to hear the true pronunciation thongõ, sometimes thonga, when listening to everyday

conversation.

/ny/ is well attested. It may be discussed if this /ny/ is found before /i/, but it is clear in most other

cases: speakers, when questioned, insist that ‘sun’ is nini (not *nyinyi), that ‘who’ is nyi (not *ni) etc. It

can be found before all oral vowels. Some /ny-/ probably come from /ngy-/, as in the case of /nyit/

‘two’ ; one case of /ngy/ is in /angyĩˀ/ ‘handle of knife’.

2.3.2. /b/ and /m/, /d/ and /n/

In many words, initial /b/ or /m/ depend on the speaker, perhaps on his or her home place. This can

affect nouns for things like ‘banana’, /musung/ or /busung/, for times like masang or basang

‘tomorrow’, mik and bik ‘to blow’, mekhe-ta or bekhe-ta ‘to the left’; or names of clans as in Mejiji or

(less common) Bejiji. However, not all names beginning in /b/ can convert in /m/, far from it. The

subject has to be investigated. Is it the trace of an older prenasalized consonant /mb/?

One important case is the (rare) m°- form of the negative prefix b°-.

The same phenomenon occurs with some words beginning with /n/: /nese/ and /dese/ ‘paddy’ (latter

is preferred), /nachhao/ and /dachhao/ (latter is preferred) ‘mouth’, /duphung/ and /nuphung/ ‘dark,

night’. In the Khiksaba Bakho (see 7.2.2), sentence 5, we hear: lu-lo ma na-na ‘do not bring bamboo

from the mountain’; na-na ‘do not bring’ is for more common da-na (defence + bring).

We wonder if some /g/ could have had a /ng/ variant; but, as we said before, the status of initial /ng/

is highly arguable. The word for ‘I’ (1st person pronoun) is gu (written Tibe. nga), and ‘five’ is khu

(written Tib. l-nga).

7 Elders and people intereted in Sherdukpen lore all agree that the true pronunciation is /thongõ/ or /thonga/.

They may discuss the final vowel, which many think should be /a/, but there is a complete agreement on /ng/,

against /ng-g/. The ‘English’ writing is the reason for the introduction of /ng-g/ among younger people.

23

2.3.3. Aspirated initials: ph, th, kh

/ph/ is often heard [f] but can be uttered in all possible ways between [f] and [ph]. /th/ is clearly distinct

from /t/, and /kh/ from /k/, /chh/ from /ch/. Literate Sherdukpen speakers never hesitate about telling

you if the consonant is to be written with or without ‘h’.

Aspirated initials are more common than non-aspirated.

Concerning initials, /w/ and /wh/ are rare but distinct from one another: whu’ ‘bird’ (‘bird’ is /fua/ in

Bugun language), whan ‘new’ (mainly under its awhan form), juwhu ‘young man’.

2.3.4. Affricates: ch, chh, j

Note about a problem with orthography (not with pronunciation).

The orthography for ch and chh can certainly be criticized, although it is customary for Hindi-speaking

people who write with the Latin alphabet8. Some Sherdukpen literate people prefer to write

respectively ts and tsh, not without good reasons because the writing ch, because of the ‘h’, suggest

an aspiration that does not exist; in the chh writing, the first ‘h’ actually means that the ‘c’ is not a ‘c’

but a /ts/, while the second ‘h’ is a true indication of aspiration – a rather stange system.

/ch/ is realized with the tongue pushed in front, while /chh/ is more palatal than dental. They may be

phonetically driven to respectively [tʃ] and [tʃh]. Hindi pronunciation habits and Tibetanized manners

(or mannerisms) are influential there. For this reason also, although a logical transcription would be

“Chering”, it seems more convenient to write Tsering.

/j/ also is between [dz] and [dʒ].

2.3.5. Consonant /h/

A good number of words begin with /h/, a frequent phoneme in the language, in all dialects (see

Lexicon); /h/ is also found between vowels, notably in colour names, ehek ‘red’, oho ‘blue’, aha ‘yellow’

– actually cases of roots beginning in /h/, with the a-prefix. A word difficult to pronounce is hẽrẽ ‘ribs’,

because the shortened first vowel brings a /hr/ cluster.

2.3.6. Final consonants

Final consonants are (see 1.1.1.) by rough order of frequency: ng, k, n, m, t, p, r. No final /l/ occurs,

although inside words some speakers ‘float’ between /r/ and /l/, for instance in labrang / lablang, see

Clusters (1.5.2.).

Syllab final /r/ appears in:

argo limestone

chandar (name of a god)

chhoskor water-mill (< Tib.)

dornok pants

dungkar conch (< Tib.)

ergo bell rope

gor cubit

8 Although the phrases ‘English letters’, ‘English alphabet’, are common in India, because most Indian people

come to know this alphabet through English, it is far better to go back to the standard designation ‘Latin

alphabet’. English people, as many other people in Europe, received it from Roman conquerors.

24

Gorbao (name of Village)

gormu a round shaped surface

jogar Plains people

jur jur heaping (in containers)

kartham culture, lore

kor hoe (cf. Tib. ‘jor)

kharbo intestine envelope

la’-nardong shin

lorjang a singers’ group

marchang rice beer with butter (< Tib.)

marso skin of some animals

ortong neck

phor fencing post

sar east (< Tib.)

Sartang the Rahung-Khoina people

Ser-pa Shergaon people

sorbo leaves fallen from trees

Srahor Doimara (place name)

sumbar cannon

ter-ba throw (with the hand)

ur-ba provoke pain

yokor spindle

yoksar neck of animals

yur-ba bring down (in bee hunting)

Yuser a quarter of Rupa

zor-pu horse with a white spot

The status of the syllable final /’/ is clear. Phonetically, it is pointed by speakers themselves with the

help a few standard examples, the favorite one being /ha/ ‘blood’ vs. /ha’/ ‘food’, where the contrast

is variously described as of height, pitch, or sheer enthusiasm; speakers who have a knowledge of tones

tend to use this description, but we doubt it really fits the facts although it is true that, as for instance

in Garo, the checking /’/ is correlated to a higher pitch. The checking /’/ is never found with another

final consonant9. It triggers the de-voicing of following suffixes, the most common being -ba (becomes

-pa after /’/) and -bo (becomes -po). The same change occurs after unvoiced finals /p/, /t/, /k/, not

after /m/, /n/, /ng/, /r/. It is easy to spot a checking /’/ in the end of verb roots, by requesting speakers

to pronounce them with a convenient suffix (if the -pa variant comes up, it is a sure sign that the final

vowel of the root is followed by /’/), while it is a more lengthy process with other parts of speech.

Consequently, on that point, the Lexicon can be considered reliable for verbs, not so much elsewhere;

in the Lexicon, verbs are written with -ba or -pa in order to stress that fact.

2.4. Syllables and word analysis

2.4.1. About words and phrases

Many syllables sound the same although they have a different meaning - such is the description by

Sherdukpens themselves. Maybe this shyness is aggravated by the loss of tones in Rupa and bigger

villages, but I guess that even with a language with tones, their idea would remain the same: they do

not feel completely at ease with isolated words, nouns or verbs or other parts os speech, and they

easily propose sentences or phrases in order to help delineating a better definition. Those who have a

9 In Miju Mishmi for instance, checked nasals (such as /’ng/) are very common.

25

feeling for grammatical precision remarked that a lexicon should include phrases; otherwise, words

would remain difficult to describe, or would be lost in a variety of possible meanings.

The answer of standard lexicography is to open a new line when the same syllable or root give a clearly

distinct meaning: we decide that this is a different word, with the same sound. There is of course a

proportion of arbitrariness in such decisions, but it is a safe procedure. Nevertheless, it is true that,

especially with monosyllabic roots of verbs, the identification of a verb without a context is close to

impossible.

2.4.2. Syllable endings

The Lexicon (here reduced to c. 1600 words10) was searched for possible syllable endings. Numbers are

given below. It should be stressed that such countings give only an approximation.

Vowels:

a e i o ö u ü ao

98 64 119 171 16 138 57 53 =716

ã ẽ ĩ õ ȫ ũ ǖ aõ

0 28 13 43 13 29 10 2 =138

To which add :

iao

4

Moreover, a number of non-standard vocalic endings are found:

ai Jowai

aĩ adu-nyaĩ, aĩ, yanlaĩ, khãi or khaĩ

ẽi ik kuẽi

iũ riũ

uõ juõ

Consonants:

-m -n -ng -r -p -t -k total

-a- 34 48 72 9 22 21 22

-e- 1 2 26 5 3 13 13

-i- 3 17 55 1 33

-o- 21 4 49 9 11 5 59

-u- 1 0 53 5 26

-ü- 1 1 30 1 14

37 40 167 244

28 28

61 72 285 418

690

10 Words or suffixes that appear several times have been counted only one or two times, depending on their

use.

26

Three diagrams follow. In the 1st one, are given the proportions of oral (V ora) and nasal ( V nas)

vowels, of voiceless (C) and nasal (C nas) consonants, and /r/ (R). In the 2nd and 3rd are shown the

details of the proportions for oral vowels and nasal vowels.

2.4.3. Syllable onsets

Vowels:

syllable endings

V ora V nas C C nas R

oral vowels

a e i o ö u ü ao

nasal vowels

aN eN iN oN öN uN üN aoN

27

a e i o ö u ü

109 2 23 20 1 8 0 =163

ã ẽ ĩ õ ȫ ũ ǖ

0 4 1 2 0 0 1 =8

Consonants:

p t k ch s

24 43 38 49 130 284

ph th kh chh

64 60 101 92 317

b d g j z

97 83 60 70 66 376

m n ng ny

95 59 1 59 214

To be added:

l r w wh y h

69 46 18 1 48 51 233

Although the partition of the 3rd category under ‘to be added’ is somewhat arbitrary, it is useful to

present a schema for them.

The ‘Types’ diagram that follows displays the proportion of these categories: oral vowels (V ora), nasal

vowels (V nas), voiceless consonants (C p), aspirated consonants (C ph), voiced consonants (C b), nasal

consonants (C m), and the rest (l, r, w and wh, y, h).

It should be noted that a vast majority (89 %) of syllabic onsets are with a consonant. This is all the

more striking since most words beginning with a vowel (approx. 114 / 171 in the Lexicon, c. 66 %)

actually have the a- prefix.

Thiese statistics are to be read with care, and interpreted with due caution: the Lexicon from which

they are derived is c. 1600 words only. Moreover, some words are repeated more than others if they

types

V ora V nas C p C ph C b C m other

28

happen to produce compounds11. For instance ik ‘hand, arm’ is found 15 times because there are 14

compounds beginning with this root; they form the major part of the 21 words beginning with /i/.

2.4.4. Mono-syllables, verbs and nouns

Most verb roots and a good number of nouns are 1-syllable. More nouns are 2-syllable, and in this

case, we either have a real compound and each syllable is pronounced, or a strong accent on the last

syllable which tends to reduce the first vowel to a short copy of the next vowel (see 1.5.2)

Examples of verb roots are:

ban- have a dream lan- look at

dan- know le’- stop, block

dap- adopt len- take over

dok- big ling- drink

dok- burn (jungle) lo- spread (seeds)

dük- put on (jacket) long- completed

düng- sit down lu- bless

jing- sleep mat- lie down

jor- attach mik- blow

ẽ’- know mok- fight

hĩ- fall down na- bring

hin- lie down nap- cover, put on shawl

hung- be afraid no- stretch

i- die nö’- be possessed

ke- give birth nyu- thirsty

ku- chew nyung- desire, want

kha- enter, go in nyü’- put on (pants)

khe’- cry o- see

khe’- squeeze, milk o’- kill

khi- stand up õ- dry

khik- put on (necklace) ö’- be in hurry

khit- éternuer ong- go

khok- put on (belt) phi- present, offer

khong- carry phing- full

khop- conceal, cover phong- sacrifice

khö’- spit ra’- prepare, cook

khung- be hungry ram- come

khyi’- quake (earth) rek- shoot (bow)

khyom- drop (smthg) reng- feel cold

There are some examples of nouns with one syllable:

ba fire hek louse

ban dream hĩ flat land, plain

bao upper ground hing wood, tree

bũ a sp. of tree hõ duck

dang vegetable hũ nature’s call

dȫ demon hũ’ salt

du’ child ik arm, hand

11 For the coounting of onsets, we included all entries of the Lexicon.

29

jang north khao door

ji’ stranger kheng horn

jo tea khi cane

jok wool khik necklace

jong grass kho: water

ju potato kho’ stick, rod

gam box, trunk khǖ garden

gö reptile ya dung

gong stem (plant) yak yak

gü wealth yam house

ha’ food, rice yo’ clothes

ha blood yuk basket (one sort)

2.5. Morpho-phonemics & consonant clusters

2.5.1. Reduced first vowel: the problem

Consider the following contrast:

gu no zilik ram-bo I feel angry

s1 Adv angry come-Cnt

gu no zi lik ram-bo I’ll go for a pee

s1-Adv urine pass come-Cnt

In the first example, the word zilik should be pronounced with its first syllable short, [zlik]. In the second

example, the word zi ‘urine’ has its own accent.

When a noun or a verb form has two identical vowels (for instance chholo ‘elder brother of father’),

the first vowel is normally strongly reduced or erased (we hear chholo or even chhlo). This give a strong

iambic turn (short syllable + long syllable) to the language – a feature which happens to be common in

South-East Asia12. Some examples with ‘full’ and ‘short’ pronunciation:

full short

baccha pchha chicken

bachham pchham snake

bazao bzao thorn

bechhẽ pchhẽ worms (in meat)

bisi psi [pʃi] four, 4

bichi’ pchi’ pine tree

bocholõ bcholõ sparrow

chökö chkö bee

chheple chhple flat, pressed flat

chhoko’ chhko’ spoon

chholo chhlo older male of same clan

chhulung chhlung hole

dochhao dochhao mouth

duphung dphung night

In the discussion that follows, we will write

- C for consonant

12 This is what Jim Matisoff calls a sesquisyllabic pattern. Sesqui- is a Latin term that means ‘a half plus one’.

30

- V full vowel

- v reduced vowel

A consequence of the strong accent on the last syllable should be (we would expect it to be) the

formation of new consonant clusters (see 1.5.2): instead of CvCV, we would have CCV. An example is

the word chhokok ‘a group of citizen paying tax together’; the word is actually pronounced chh°kok or

chhkok (CvCVC or CCVC). Why do we write chholo and chhokok, instead of chhlo and chhkok? Because

in slow speech, the first vowel is heard, although shorter than the second one. The degree of

pronunciation of first syllables in slow pronounciation heavily depends on the speaker: some know

that chhlo (CCV) is actually chholo (CvCV), some do not. Most people do know, and would admit that

in all cases the first syllable is shorter. After a few generations we would expect numerous words like

chhlo (CCV) or like chhkok (CCVC), with no trace of the possible chholo or chhokok. Yet, very few

consonant ‘solid’ clusters appear in the present-day language, and the speakers still consider those

words to be with two syllables.

Real consonant clusters at the beginning of words do exist. They are typically formed with /s/ +

consonant, such as spu (CCV) ‘cow, cattle’ or stu (CCV) ‘horse’: no speaker would think of them as

*supu (CvCV) and *sutu and they are never pronounced in this way. Comparative linguistics shows that

in the past they indeed were *supu and *sutu13, but since nobody ever pronounces them that way

even in slow speech, it seems quite inappropriate to write them with two syllables. However, the

degree of pronunciation of those first syllables also depends on phonetic factors. Clusters CCV are

more easily pronounced CvCV before /r/ and /l/: sri ‘sacred stone’ is siri in slow speech, and slẽ is also

given as sẽlẽ ‘spleen’. When /s/is followed by a /p/, /t/, /k/ consonant, we usually have a “true” CCV

word.

Since this phenomenon is observed only when the two vowels are identical, it is quite possible to

describe the facts in a different way. We may consider that the first syllable is reduced and has no

vowel of its own, only an echo from the next vowel. In this alternative presentation, we should write

chh°kok ‘taxation group’ or s°lüng ‘heart’, with a small sign ‘°’ indicating an echo vowel, whose actual

sound depends entirely upon the next true vowel. This would be more accurate, but more difficult to

read. We retained the description with ‘°’ only in grammatical explanations.

2.5.2. Short vowels in grammatical prefixes

Verb roots are mostly mono-syllabic (see 1.4.4.), and verb forms are usually made with prefixes or

suffixes attached before or after the verb root. Negative markers are prefixed, either in imperative or

in declarative sentences (see 2.8. and 2.9.). The result of Negation + Verb root is a two-syllable

derivative14, for instance ling- ‘drink’ and bi-ling- ‘not drink’. The stress or accent is then on the second

syllable (the verb root), and the first syllable (the negative prefix) is shortened to the point of having

only a reduced vowel (biling-15) or no vowel (bling-). This vowel is heard as a shortened copy of the

next full vowel, viz. the vowel of the verb root (See 1.5.1.) Since it is only an ‘echo-vowel’, the negative

prefix b- does not have its own vowel. But since, it has a short vowel nevertheless, the ‘colour’ of which

13 In both these cases, the first syllable su- come from a root meaning ‘animal’. A ‘memory’ of this root in

Sherdukpen is found in su’ ‘meat’. 14 It is useful here to make a distinction between compounding and derivation. A compound (the result of

compounding) is build with two words of about the same status, for instance phudo associating two kinds of

deities, phu and do. A derivative, rsulting from derivation, is built of a root and an affix.

15 We write the first /i/ smaller, in order to suggest it is shorter or quicker.

31

will depend on the next vowel, we have to write this prefix with a small sign indication the ‘echo vowel’.

Here I use a small circle: b°-.

In the same way, the shortened vowel with the negative imperative prefix d°- is represented in this

book by a small circle. For instance the negative imperative of khe- ‘cry’, d°-khe-is heard as dekhe- with

a very short /e/ in the first syllable because the root verb is with a vowel /e/. The same form with the

verb lan- ‘look’ is d°-lan- and sounds as dalan! ‘don’t look’, with a short /a/. The reason for this device,

is that we do hear a vowel, although a short one, with these prefixes; but we cannot say which one

until we know with which verb the prefix comes.

2.5.3. Short vowels in nouns

For the noun prefixes b°- and g°-, see 2.3.1. In the previous cases, the phonetic shortening of the first

syllable is connected with its grammatical status: the first syllable is not a word with a meaning, only a

prefix. This is distinct from compounding, where the two parts of the compound have or had a

semantic value and are sounded, for instance with lo-blang ‘deity residence’. Other examples are

khomi ‘citizen’ (etymologically ‘tax-payer’, from kho ‘tax’ and mi ‘person’), khodȫ ‘crocodile’ from kho

‘water’ and dȫ ‘monster, demon’. In such compounds, the speakers are often able to discern and

explain each element.When a noun is written with two different vowels, this is a sign that the first one

is not shortened16. Yet, this is further complicated by the “semi-prefix” status of some frequently used

roots.

A nice example is the ‘paddy family’: nese ‘paddy’, nichhi ‘husk’, nodop ‘husked rice’. It is clear that the

three nouns are compounded with a n°- root broadly meaning ‘paddy’, the vowel of which is so

reduced as to be only an echo of the next one. The same conclusion would be arrived at, if we were to

consider the numerous nouns denoting ‘big animals’ and beginning with s°-, a prefix that comparative

Tibeto-Burmese linguistics shows related to or identical with su ‘meat’17.

sdung monkey

ska sheep

skan boar

ski deer

sobo, sbo porcupine

spu cow, cattle

stong elephant

stop mouse, rat

stu horse

stung bear

sumu, smu mithun, buffalo

suwo, suo ox, bull

In both cases, a syllable that has or had the full status of a noun, with sound and meaning, developed

into something like a prefix, both for the sound (reduced to a copy vowel) and the meaning (expanded

to a category).

2.5.4. Names for numbers in 11 to 17.

See 4.1.

16 Some words, usually ‘ideophones’ (words, for instance, where repetition suggests a stronger meaning) have

a double syllable, both full: susu ‘different’. Such is the case in baby talk (see: 4.6.). 17 Perhaps should we also think of suak ‘pig’, sao ‘hyena’, se ‘rhino’.

32

10 sõ

11 sahan

12 sinyik [sinyik]

13 suũ [suũ]

14 sambisi

15 sangkhu

16 sangkhüt, saĩkhüt

17 sangsit, saĩsit

18 sõ sarjat

19 sõ dekhü

20 khan, khaĩ

This list in Standard Rupa speech shows (1) that the word for ‘ten’ is sõ, (2) that this word is

transformed when compounded with unicts. For 11, 12, 13, ‘ten’ is reduced to s°- as the copy vowel

shows; from 14 to 17, it seems that a /san/ basis operates, although the /n/ is adapted to the following

consonant; for 18 and 19, sõ reappears.

A similar but weaker adaptation occurs in names for days and years (see 4.2.), to a point that depends

on the speaker: some speakers do assimilate and say lingkhit, some do not and say lin khit.

2.5.5. Hunting for clusters

In many cases, the first vowel is strongly reduced, and sounds as a short copy of the next vowel. So,

sumu ‘mithun, wild cattle’, is often pronounced sumu with a reduced first vowel, or smu. In this way,

pseudo consonant clusters are formed, sometimes difficult to discern from the true ones. Spu ‘cattle’

or stu ‘horse’ are never pronounced with a first short vowel (*supu ou *sutu), although on comparative

grouds the su- in these words is quite comparable to the su- in sumu ‘wild cattle’. It is clear that the

quality of the micro-vowel depends on the character of the consonant that follows, and that it is easier

to hear the first vowel in sumu because of /m/ than in stu, because of /t/. For the same phonetic

reasons, it is easier to perceive the micro-vowel bulu ‘council’. If this is true, there is no phonemic

distinction between the first (spu, stu) and the second (smu, blu) pair of words, then no grammatical

feature to separate them.

It is often possible to show that the reduced first vowel, that has now become a copy of the next vowel,

was in earlier times a vowel in itself. For instance in a number of compounds the first syllable of which

is a-. In the following names of colours: aha ‘yellow’, ehek ‘red’, oho’blue’, it is clear that the first vowel

is the a- prefix of ‘adjectives’, now modified, like in abẽ ‘other’ or ablo ‘tasty’. The noun ihĩ ‘root, nerve’

is probably from a+hĩ, but good speakers insist that the first vowel is /i/, not /ĩ/; uhu ‘prayer’ may be

from a+hu. In all these cases, the consonant is /h/, but the same vowel reduction occurs with other

consonants, as in ayung ‘shadow, soul’ often pronounced oyung or uyung, or ojo ‘high’ which certainly

is from ajo.

In words like khla-po ‘go-between’ or khleng-thong (the name of a sub-clan), it is difficult to state if

khla and khleng are from older *khala and *kheleng, and impossible to see if the first syllable had a

distinct vowel in the past. The word besme ‘winter migration’ is probably to be analysed be+seme since

no Sherdukpen syllable ends in /s/. The word kamrang ‘anyway’ probably belongs to the –rang group

of words with distributive meaning (see mu-rang, nyi-rang), but it is difficult to know if it is kam+rang

or ka+ma+rang.

On a list of 73 cases of such clusters, 31 have a second component /r/, 24 have a second component

/l/. We give a few examples of those two types, then the remaining cases.

33

b r brop hearth

j r jiring [dʒring] people, man

h r herẽ [hrẽ] rib(s)

m r mi-ring [mring] sister

s r siri [sri] sacred stone

b l bulu [blu] council, community

ch l buchulung [pchlung] cockroach

kh l khala-po [khlapo] middleman

m l malang [mlang] fruit

s l sẽlẽ [slẽ] spleen

Other cases

s b sobo [sbo, zbo] porcupine

g d godong [gdong] pit (natural pit in soil)

s d ik-sdop finger ring

s d sodop [sdop, zdop] paddy (husked) var. nodop.

th g thogo [thgo] here (see: thi)

ch k chokö [chkö] bee, living in tree-holes

s m sumu [smu] mithun

kh n khini [khni] auntie

kh n khunu [khnu, khnũ] Aka or Miji people

ph n phenẽ [phnẽ] itching, irritation

b s bisi [pʃi] four, 4

d s tese [che], dese paddy, unhusked

s t sütü [stü] cloud (sinti SH)

b th batha [ptha] dog

b ch bacha [p°cha] hen

b ch bacham [pcham] snake

b z bizik [bzik] flea

g z gazang [g°zang] hair (on head)

2.6. Morphophonemic smaller points

2.6.1. Phonetic variation in a few roots

The grammatical suffix paradoxically changes the word that it follows in the case of some personal and

demonstrative pronouns. See 2.2.1. and 2.2.3. for details. As far as the personnal pronouns are

concerned, the results differs according to dialect, but the principle holds good: s1 gu and s2 nang are

mostly affected. This trend is found in numerous languages. Here is a sentence displaying the same

variation in s3 wa, a less common occurrence.

() wo-’o gu-ni an-do-chhi-baũ he was told by me

s3-A s1-O tell-NGp-Fac-Pfp

For thi ‘this’ (perhaps thü is the basic form), the range of variation is wide. The chart gives an idea.

expected form actual forms

this one *thi-wa thu-wa, thwa, tha

here *thi-go thügo, thogo, thgo, thkho

2.6.2. Phonetic variation in a few suffixes

A small number of grammatical postponed morphemes show a phonetically triggered variation. See

2.5.3. Thereare small and wide variations.

34

A (unique?) case of wide variation is the locative suffix. Although the degree of assimilation depends

on the speaker, one may hear:

-ko after unvoiced stops (k, t, p etc.)

-kho after aspirated stops (kh, th, ph), e.g. thkho < thü-go.

-go after vowels and /r/, /l/

-ngo after nasals (n, m, ng)

A few cases of small (binary) variation are found:

ba / pa /pa/ only after unvoiced stops (-p, -t, -k), and glottal /’/ included;

/ba/ after nasals (-m, -n, -ng) and -r.

The same is true with –bo / -po.

A few other suffixes show a variation, but it does not seem to be conditioned by phonetics, only by

rapid speech.

2.6.3. han > aĩ

One frequent but puzzling realization is /ãi/. The intriguing fact is that no phoneme /ã/ seems to be

functional in the language (see 1.2.2). We hear [ã] when there is /a/ in nasal contexts, but no phonemic

status can be assigned to it. The solution to the small enigma is that /ãi/ is a ‘modern’ pronunciation

of /aĩ/. This diphthong comes from han ‘one’:

nuphu haĩ, nuphu nyik jao-na-ma he can stay 1 or 2 nights.

night one, night 2 stay-Gp-Ipf

See the description of indefinite pronouns in 2.2.5.

2.6.4. Other notes

The Defense form of na- ‘bring’ usually is na-na (Not *da-na). See 1.3.2.

Verb roots ending in /m/ often exhibit a [n] variation when before a alveolar consonant. With ‘to

come’, one says ram-ba but ran-la; with ‘to tell’, one hears am-baũ ‘he said’ but an-do-baũ ‘he just

said’.

2.7. Pronunciation and society

Dialects have been touched upon in the introduction and will be briefly discussed later, but it should

be noted in this chapter about phonetics and phonemics that older people often have /ü/ and /ö/

where younger ones have /i/ and /e/ or /ẽ/, respectively; and that the elders have /ǖ/ and /ȫ/ where

younger people have /ĩ/ and /ẽ/.

elder e ẽ i ĩ ö ȫ ü ǖ

younger e ẽ i ĩ e ẽ i ĩ

So, where elders have 8 distinct phonemes, younger people often have 4 only. Younger people often

agree, with variations in shame, that they simplify the ‘more beautiful’ pronunciation of their elders.

In this book, we follow the ‘older pronunciation’18. The role of Hindi phonemics in this ‘simplification’

is obvious. Moreover, most foreign people with whom we could talk (Bugun for instance, whose

language is not phonemically easy) admit that Sherdukpen pronunciation is difficult.

Another but quite different problem is the Tibetanizing fashion. Among the previous generations (see

7.3.), say until the Chinese Agression (1962), Tibetan was a fashionable language and being able to

18 The problem we describe here is different from the existence of tones in the language of some older people.

35

speak and to write it was a classic sign of the elite. After the border was closed, things changed; yet

Tibetan folklore and lore remained somehow influential and, also through Monpa regions

(conspicuously in Shergaon), the pronunciation of names is gladly ‘tibetanized’. The name of the hero

Asu Japtong can be heard either as [japtong], [dȝaptong] or [dȝyaptong] and [gyaptong], and some

people told us one should say [gyapteng]. Names of characters or roles in performances are a possible

field for exploring this topic.

2.8. Quick speech

We are sorry to offer here only a few remarks for a topic that seems extraordinarily important. Because

Sherdukpen language is, most of the time, easy to reduce to monosyllabic units, we can often

understand which stuff a compounded word is made of, and we can, at the same time, realize how

these units are modified when the compound is pronounced ‘rapidly’ - and that means normally. Let

us try to make the point very clear. When you describe (or write) a monosyllabic language or a language

where each syllable maintains its shape and status, writing consists in putting each syllable in its right

place, in its turn. Now, Sherdukpen is not like that. Many words are made of a number of syllables,

sometimes as many as 4 or 5. Normally, each syllable is useful in its own way, as a kind of brick in the

building, and has to be identified clearly. But in some cases, the compound makes a unit in itself, and

neither the speaker nor the listener has to analyze the components in order to understand it. This is

especially true with pronouns (personal, demonstrative, interrogative, indefinite). Although it is quite

possible to discern the different elements which build up such pronouns when they are pronounced

in slow speech (a rather formal experience, of course, but a possible one), yet in normal speech the

multisyllabic pronouns are ‘taken for granted’ as such, pronounced in one move, and the syllabic

borders of the compounding units disappear.

This is a rather common process. In written languages for which we have old documents that show

(more or less) how the language was pronounced two or three centuries before, we can see how some

words ‘imploded’ so to speak, and how a strong stress on one syllable shortened the other syllables,

so that a word which, some centuries before had 4 syllables, now has only two or one. For instance

the English word captain, two syllables, comes from a Latin word capitāneus which had 4 or 5. What

happened? This was a foreign word, borrowed because of military practice; it had a strong accent on

the long ā in the middle, and English usage made that syllable still longer, while it eliminated those just

before and just after; the first syllable, although less prominent, remained and the word has now two

syllables. The main point is that we know the Latin word, and some other stops in that journey through

forms and time, because they were written in documents. But the hundreds of languages in the

Himalayas, with few exceptions like Tibetan, Kashmiri, Lepcha in Sikkim, Ahom in Assam, Meithei in

Manipur, Burmese of course, are usually not written and for the vast majority of them we do not know

how the language was spoken in the past. It would seem, then, that the present stage of a language

like Sherdukpen is some snapshot that we can take when we listen to it now, record it, study it, write

it.

Yet, the truth is very different. A language is not as solid as a statue or a stone monument. It is always

diverse not only because it is spoken by different people, older and younger, urban and village-folks,

grandmothers and young students, but also because each speaker has different ways of using the

language. When friends talk to us, they take care of pronouncing carefully, even without noticing it

sometimes; but suppose a Sherdukpen friend comes in, then the rhythm of the conversation will

change very much, words will become swift and short, and a kind of artistry with speed will be

appreciated. What is important for us here, is that formal speech (we mean public formal speech by

elders, by gaonburas or important members, or by Asu dochhao during Khiksaba festival) is slow. It is

not only full of flowery expressions, it is also dramatically slow (and low, sometimes difficult to hear!).

36

Because of this tradition maybe, or some other reasons as well, Sherdukpen elder people can speak

both ways: the slow speech, and the quick normal speech. When you ask a question, they will slow

down and explain, because slow speech maintained very clearly the component units of the words.

But in in the normal flow, you would hardly hear them, just the synthesized result.

Quick speech may give us an idea of what will be the language soon. There are children that only know

that quick shortened speech. But slow speech, formal speech, maintains the old building bricks in their

own place, and Sherdukpen people can usually tell you, by intentionally slowing down their speech,

what the units are in words. This is not exactly being bilingual; it is something different, because slow

speech and quick speech ARE the same language. Culture is most aptly described, perhaps a bit

provocatively: having several cultures. When you are not the child of only one world, when you are

able to look at your birth-place with love and with distance, when you can play with the instruments

of intelligence (which are words, phrases, sentences, puns, jokes, careful allusions in a speech), then,

this is culture.